selective attention determines emotional responses to novel visual stimuli

Selective Attention Determines Emotional Responses to Novel Visual StimuliAuthor(s): Jane E. Raymond, Mark J. Fenske and Nader T. TavassoliSource: Psychological Science, Vol. 14, No. 6 (Nov., 2003), pp. 537-542Published by: Sage Publications, Inc. on behalf of the Association for Psychological ScienceStable URL: http://www.jstor.org/stable/40063909 .

Accessed: 15/06/2014 08:32

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Sage Publications, Inc. and Association for Psychological Science are collaborating with JSTOR to digitize,preserve and extend access to Psychological Science.

http://www.jstor.org

This content downloaded from 185.44.79.85 on Sun, 15 Jun 2014 08:32:46 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=sage

http://www.jstor.org/action/showPublisher?publisherCode=assocpsychsci

http://www.jstor.org/stable/40063909?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


SELECTIVE ATTENTION DETERMINES EMOTIONAL RESPONSES TO NOVEL VISUAL STIMULI

Jane E. Raymond,1 Mark J. Fenske,1 and Nader T. Tavassoli2 'University of Wales, Bangor, Wales, United Kingdom, and 2 London Business School, London, England,

United Kingdom

PSYCHOLOGICAL SCIENCE

Research Article

Abstract - Distinct complex brain systems support selective attention and emotion, but connections between them suggest that human behavior should reflect reciprocal interactions of these systems. Al- though there is ample evidence that emotional stimuli modulate attentional processes, it is not known whether attention influences emotional behavior. Here we show that evaluation of the emotional tone (cheery/dreary) of complex but meaningless visual patterns can be modulated by the prior attentional state (attending vs. ignoring) used to process each pattern in a visual selection task. Previously ignored patterns were evaluated more negatively than either previously attended or novel patterns. Furthermore, this emotional devaluation of distracting stimuli was robust across different emotional contexts and response scales. Finding that negative affective responses are specifically generated for ignored stimuli points to a new functional role

for attention and elaborates the link between attention and emotion. This finding also casts doubt on the conventional marketing wisdom that any exposure is good exposure.

An important aspect of human behavior is selection. In many different settings, from art galleries to supermarkets, humans select specific objects to be the targets of their actions, while ignoring or

avoiding other objects. Inherent in this activity are two different but related processes. One is a strategic, object-oriented perceptual process (Duncan, 1984) that determines what and where objects are and involves the brain's selective attention system, which facilitates processing of specific task-relevant stimuli and suppresses processing of

distracting stimuli (Kastner & Ungerleider, 2000). The other is an emotional process that evaluates object representations in terms of current and future goals (Ortony, Clore, & Collins, 1988). To smoothly guide action toward or away from specific objects, coordination between selective attention and emotional systems is necessary. Such coordination is probably supported by abundant connections between the fronto-parietal cortical network subserving spatial attention

(Hopfinger, Buonocore, & Mangun, 2000) and the complex network

mediating emotion (Le Doux, 1996; Panksepp, 1998; Rolls, 1999). Recent human neuroimaging studies have shown that brain structures common to the two systems (e.g., the anterior cingulate and orbito- frontal cortices) are selectively activated during selective attention and emotional evaluation tasks (Armony & Dolan, 2002; Bush, Luu, & Posner, 2000; Vuilleumier, Armony, Driver, & Dolan, 2001). These

findings suggest that human behavior should reflect reciprocal interactions of these two major brain systems.

Indeed, there is now ample evidence that emotional stimuli pro- duce specific effects in selective attention tasks. For example, images

of angry faces attract (Eastwood, Smilek, & Merikle, 2001; Vuil- leumier & Schwartz, 2001) and hold (Fox, Russo, Bowles, & Dutton, 2001) attention more than emotionally neutral or novel stimuli. Simi- larly, aversive words are more easily detected than neutral words in an attentional blink task requiring rapid successive attentional processing (Anderson & Phelps, 2001; Raymond, Shapiro, & Arnell, 1992). However, evidence for the influence of attentional state on emotional behavior is lacking. Although many studies have reported that emotional evaluation of stimuli can be manipulated by repetition (Born- stein, 1989; Zajonc, 2001), nonfigural stimulus properties such as brightness (Reber, Winkielman, & Schwarz, 2000), and prior presentation of other emotionally salient stimuli (Murphy & Zajonc, 1993), no study has systematically manipulated the observer's attentional state so that the influence of attention on emotional responses could be observed.

To determine how selective attention might affect subsequent emotional evaluations, we combined a simple visual localization task with an emotional evaluation task. (See Fig. 1.) For the localization task, two complex novel stimuli, each drawn from one of four distinct types of patterns and without any obvious emotional valence, were briefly presented: one on the right and one on the left. Participants were required to locate (left or right) a previously designated target type as quickly as possible. After this selective attention task, a single image (the test stimulus) was presented, and participants evaluated its emotional tone using a simple 3-point scale. In the experimental conditions, the test image was one of the previously viewed stimuli, either the target or the distractor. In the baseline condition, the test image was novel and was always a different type than either stimulus presented in that trial's localization task. Comparing affective ratings for test images that were just previously attended or ignored, or were never seen before, allowed us to determine whether the attentional state active during initial exposure to a stimulus had consequences for its subsequent affective evaluation.

In the first experiment, one group judged test images as "cheery" or not; another group, viewing the same stimuli, judged images as "dreary" or not. Using opposite response scales allowed us to examine whether attention specifically modulates emotional responses or whether it acts more generally by biasing all judgments in the same way. In this experiment, we also controlled the emotional context of each trial by first briefly presenting either an emotionally valenced or a neutral prime. Positive and negative emotional primes are known to modulate affective evaluation of subsequently presented stimuli (Mur- phy & Zajonc, 1993). In this experiment, these primes served to test that evaluations of stimuli made later in the trial were predictably mal- leable by affective manipulations and also to determine whether emotional context interacts with attentional state. In the second experiment, we replicated the finding of the first (without the use of emotional primes) and tested whether the effects of attention were specific to stimuli just previously viewed in the localization task or whether they

Address correspondence to Jane E. Raymond, School of Psychology, University of Wales, Bangor, Gwynedd LL57 2AS, United Kingdom; e-mail:

[email protected].

VOL. 14, NO. 6, NOVEMBER 2003 Copyright © 2003 American Psychological Society 537




Selective Attention Determines Emotional Responses

Fig. 1. Types of stimuli (a, b) and the sequence of events on an experimental trial (c). Affective primes (a) displayed positive or negative affect, or were scrambled to be affectively neutral. Circles and squares patterns were first used in a simple visual search task (Tl) and later rated on a 3-point scale (T2); examples are shown in (b).

generalized to other, novel stimuli of the same pattern types as the stimuli previously attended or ignored.

To anticipate, we found robust effects of attention on the subsequent emotional evaluation of stimuli. Just previously ignored stimuli (and others of the same type) are devalued relative to just previously attended stimuli (and others like them). Moreover, ignored stimuli are devalued compared with novel stimuli that are distinct from previously attended or ignored images.

EXPERIMENT 1 The main goal in our first experiment was to determine whether at-

tention during initial exposure to a novel stimulus modulates subsequent affective responses to that stimulus or, more generally, biases

judgment of any stimulus dimension. Our strategy was to have two different groups of participants use two different scales with opposite affective valence. If attention modulated judgments on stimulus at- tributes generally, then the two groups would be expected to use the response scales similarly; if attention specifically modulates affective evaluations, then oppositely valenced response scales would be used in opposite ways. To obtain additional evidence that we were measur- ing affective responses in this experiment, we used affective primes at the beginning of each trial. Because these are known to modulate emotional responses, we expected them to exert a modulating effect if our response measure was indeed tapping an emotional process. Moreover we were interested to know whether affective priming effects would change with, or be independent of, attentional modulation.

Method

Participants

Eighty students and members of the University of Wales, Bangor, community (57 females and 23 males; mean age = 24.2 years) partic-

ipated in exchange for course credit or money. Half of the observers

(randomly assigned) judged images as "cheery" or not, and the other half judged images as "dreary" or not. Participants in both this and the second experiment were naive to the exact aims of the experiment, had not participated in a similar experiment, and gave informed consent

prior to participation.

Apparatus and stimuli

The experiment was conducted on two Macintosh G4 computers with 55.9-cm Mitsubishi monitors (color setting: "thousands"; resolu- tion of 1 152 X 870; refresh rate = 85 Hz) running PsyScope experimental software (Cohen, MacWhinney, Flatt, & Provost, 1993). Stimuli were presented on a uniform white field at an average viewing distance of 70 cm. Single stimuli were presented at the center of the

display. Pairs of stimuli were centered approximately 4.3° to the left and right of the display center. All alphanumeric stimuli (i.e., "+" and "?") were presented in black 48-point Times font.

Examples of the stimuli are shown in Figure 1 . Affective primes (see Fig. 1 a) were schematic faces displaying positive or negative affect, or a control stimulus with the same low-level features scrambled to make it affectively neutral. The affective primes were drawn in black lines, and each subtended approximately 4.3° in diameter.

The novel complex visual patterns were generated by randomly varying the elements that defined each as being one of the four pattern types. Two of the pattern types were used as test images (see Fig. lb): "squares," created by randomly (with replacement) selecting the color of each section of a 5 X 5 grid from a 10-color palette; and "circles," created by randomly varying the x,y coordinates and radius of 20

uniquely colored circles on a gray background. The other pattern types were used only in blocks of baseline trials: "polygon," created by ran-

domly varying the x,y coordinates of 10 vertices of the angles forming a red-filled irregular polygon on a gray background; and "squiggles," created by randomly varying the xyy coordinates of 10 points con- nected by a single red smoothed line on a gray background. Each pattern subtended approximately 3.3° in height and width.

Procedure and design

The general procedure and sequence of events for each trial is shown in Figure lc. At the start of each trial, an affective prime dis-

play was presented for 200 ms. It consisted of two matched schematic face stimuli, presented to the left and right of a central fixation cross. After a brief (50-ms) interval, the display for the localization task was

presented for 100 ms and then replaced by a blank screen. The display for the localization task also consisted of a pair of stimuli positioned to the left and right of fixation, but in this case, two different stimuli were used. In the no-exposure baseline trials, these stimuli were a

polygon pattern paired with a squiggle pattern, and in experimental trials, they were a squares pattern paired with a circles pattern. At the

beginning of each block of trials, one of the patterns was designated as the target category, and participants were told to locate the image of that type as quickly as possible using adjacent keys that were labeled "<- " and "-*," respectively. After a localization response was made, a

question-mark prompt was presented for 1 ,000 ms, and then a single test stimulus was presented for 500 ms. This image was always either a circles or a squares image.

Test images were preceded an equal number of times by each type of prime (positive, negative, scrambled) and were, for each prime con-

538 VOL. 14, NO. 6, NOVEMBER 2003




Jane E. Raymond, Mark J. Fenske, and Nader T. Tavassoli

dition, equally often the same image that previously appeared as the target, the same image that previously appeared as the distractor, or a novel image unlike either seen in the prior localization display. Half of the observers judged test images as cheery or not, and the other half judged the same images as being dreary or not. The two groups used the same three response keys, which were arranged in a horizontal row and labeled "not at all," "maybe," and "definitely," going from left to right.

Each observer was tested with two baseline blocks (30 trials/block) and two experimental blocks (60 trials/block). The order in which each image type served as the target in the localization task was coun- terbalanced across groups. Half of each group selected circles in the first experimental block (ignoring squares) and selected squares in the second experimental block (ignoring circles). The remaining observers used the opposite order (i.e., selected squares in the first block and circles in the second block). In this way, all observers rated the same test stimuli and had equal exposure to the same localization-task stimuli; they differed only in how they used each type of stimuli to perform the localization task.

Data analysis and response-scale conversion

In this and the following experiment, only scores for trials in which targets were correctly localized were analyzed. This criterion excluded only 3% of trials, as localization accuracy was consistently high across response scales and primes. Rating responses on the two scales were then converted into numerical values in such a way that the same emotional tone was reflected by the same numerical value, with smaller values reflecting more negative evaluations and larger values reflecting more positive evaluations. Specifically, responses using the cheery scale were assigned a value of 1 for "not at all cheerful," 2 for "maybe cheerful," and 3 for "definitely cheerful." Conversely, responses using -the dreary scale were assigned a value of 3 for "not at all dreary," 2 for "maybe dreary," and 1 for "definitely dreary." If observers expressed similar emotional responses, despite using oppositely valenced responses scales, then there should be no difference between the groups in their converted rating scores. All further references to rating values refer to these converted values.

Mean baseline ratings for each observer were submitted to a mixed-factors analysis of variance (ANOVA). Because block (first vs. last) did not interact with either response scale (cheery vs. dreary; F < 1) or prime (positive vs. negative vs. scrambled; F < 1), we averaged scores from both blocks for each condition. The resulting baseline scores were submitted, along with the corresponding experimental- trial scores, to a mixed-factors ANOVA to assess the combined effects of response scale, prime valence, and previous attention state (ignored vs. attended vs. baseline). In addition, separate mixed-factors ANOVAs were conducted to assess the effects of primes on baseline and experimental trials. Where applicable, simple effects were examined using paired-samples t tests.

Results and Discussion

The important and novel finding of this experiment (see Fig. 2) is that emotional evaluations were affected by attention state, as the main effect of attention was significant, F(2, 156) = 5A5rp < .01. Previ- ously ignored images were judged more negatively than previously attended images, t{19) = -2.67, p < .01, and more negatively than novel images, t(19) = -3.02, p < .01. The latter effect is important

Fig. 2. Results from Experiment 1 . The graph shows, for each attention condition, the rating of the emotional tone of images averaged across affective-prime conditions, response scales, and observers. The horizontal gray bar indicates the mean rating for novel stimuli; its width and the vertical lines represent ± 1 SE.

because it shows that the devaluation of previously ignored items is not simply a property of their status as nontarget stimuli. In the baseline trials, test stimuli were always nontargets but were never previously ignored. Because they were rated more positively than the

nontarget, distracting stimuli used in the selective attention task in the

experimental trials, it can be concluded that active ignoring produced the affective devaluation. Images that were previously attended in the

experimental trials were judged about the same as the novel, nontarget images in the baseline trials, *(79) = -0.35, p > .1, a finding that further indicates that it is active prior ignoring, rather than perceiving an

image as a nontarget, that lowers ratings. Although targets and distrac- tors were preexposed for the same brief duration, emotional judgments about them were significantly different according to their task relevance in the preceding localization task. This difference indicates that attentional state during initial exposure to a stimulus modulates

subsequent emotional responses to it. Neither the main effect of response scale, F(l, 78) = 2.52, p > .1,

nor any of the interaction effects involving response scale (F < 1 for each), was significant. Note that because we numerically converted rat-

ings from the two oppositely valenced scales to reflect emotional tone (see the Method section), this finding allows us to conclude that it was the emotional evaluation of stimuli that was specifically modulated in this experiment, and that the results do not reflect a general biasing of

VOL. 14, NO. 6, NOVEMBER 2003 539





responses for all judgments. For example, participants using the dreary scale were more likely to choose the "definitely" response for prior dis- tractors than prior targets, whereas participants using the "cheery" scale were less likely to choose this option for the same conditions. If any judgment task, emotional or not, were open to attentional biases, then the likelihood of selecting the "definitely" option should have been the same for a given attentional condition across the two groups of observers. That this was not the case provides clear evidence that it was the emotional tone that was affected by selective attention.

Another interesting finding is that the emotional devaluation of dis- tractors was robust across emotional context, as manipulated by the affective primes. In the no-exposure baseline trials, we found, as expected, that the affective primes significantly influenced subsequent evaluations of novel stimuli, F(2, 156) = 5.07, p < .01, and, thus, were effective at modulating the emotional context of the trials. Posi- tive primes produced significantly more positive evaluations (2.15) than negative primes (2.07), f(79) = 3.18, p < .01. This affective priming effect was not influenced by the response-scale valence; neither the main effect of response scale nor the interaction involving response scale was significant, F(l, 78) = 2.59, p > .1, and F < 1, respectively. However, in contrast to the baseline condition, the experimental trials showed that the valence of the primes made no difference for the evaluation of either previously attended images, r(79) =

-1.06, p > .1, or ignored images, f(79) = -1.45, p > .1. For each prime type, previously ignored images were evaluated more negatively than previously attended stimuli - positive primes: 2.01 versus 2.07, t(19) = -2.50, p < .05; scrambled primes: 2.05 versus 2.13, r(79) = -3.32, p < .01; negative primes: 2.05 versus 2.10, f(79) =

-2.09,/? < .05. This finding provides evidence that selective attention acts directly on emotional evaluation of subsequently viewed stimuli. It indicates that attention effects are not modulated by the effects of the ambient affective context; rather, the attention effect is robust across different contexts.

EXPERIMENT 2

In the experimental trials of Experiment 1 , the test stimuli were always identical to either the target or the distractor just seen in the prior localization task; in the baseline trials, the test stimuli bore no similar- ity to either pattern just previously seen. This leaves open the question of whether the devaluation of previously ignored stimuli occurs only for images that are exactly the same as a prior distractor (i.e., the same token), or generalizes to other images of the same general type as a prior distractor. We explored this issue in Experiment 2. Because attentional selection of the target in the localization task was based on identifying its type, it is likely that the negative affect produced from ignoring would apply to other novel images of the same type and not be confined to the same token (e.g., Tipper & Driver, 1988).

There were three important differences between this and the previous experiment. First, because attention effects were found to be independent of affective priming in the first experiment, no primes were used in Experiment 2. (A replication of the main effect would then serve to underscore this point.) Second, the target type was randomly assigned and cued prior to each trial (as opposed to be being tested in blocks, as in the previous experiment). Third, the test stimuli were either identical to one of the stimuli seen in the immediately preceding localization task (replicating the previous experiment) or were novel but of the same type as either the previous target or the previous distractor.

Method

Participants

Sixty-six students and members of the University of Wales, Bangor, community (42 females and 24 males; mean age = 22.6

years) participated in exchange for course credit or money.

Procedure

The apparatus and stimuli were the same as those used in Experi- ment 1. The trial sequence was also the same except, instead of

primes, a single word was presented to indicate to the observer which stimulus type to locate in the upcoming localization task. The target type (squares or circles) and its subsequent location in the display (left or right) were randomly varied on successive trials with the constraint that the two pattern types both served as target and appeared in each location equally often. Target localization and affective evaluation re-

sponses were entered by the participant in the same way as before. All

participants rated test stimuli for cheeriness, and test stimuli were al-

ways squares images. The experimental session consisted of two blocks of 24 trials each: In one block (same-token), all the test stimuli were exactly the same as the previously ignored or previously attended stimuli (replicating conditions in Experiment 1); in the other block (same-type), all the test stimuli were of the same type as the previously attended or previously ignored stimulus but were novel exem-

plars. Half the participants completed the same-token block first and half completed the same-type block first.

Results and Discussion

Figure 3 shows the mean ratings obtained in Experiment 2 for test

images that were either the same token or the same type as previously attended or ignored items. (The data from Experiment 1 are replotted in the figure for ease of comparison.) For both same-type and same- token conditions, mean ratings for previously ignored stimuli were lower (by the same amount) than mean ratings for previously attended stimuli. Both the magnitude of the attention effect and the mean rating values obtained in each condition closely replicate the results of Ex-

periment 1 . A 2 (attention: attended, ignored) X 2 (test stimulus: same

type, same token) repeated measures ANOVA on these ratings con- firmed that, as in Experiment 1 , the main effect of attention was significant, F(l, 65) = 9.33, MSE = 0.035, p < .01. Attention did not interact with test stimulus (F < 1). Although same-token ratings were slightly higher than same-type ratings, overall, the main effect of test stimulus was not significant, F(l, 65) = 1.82, MSE = 0.91,/? > .1.

The results of this experiment serve two purposes. First, the novel

finding from the previous experiment is clearly replicated, confirming that attention can reliably modulate affective responses. This replication was obtained without using prime stimuli and therefore under- scores our previous claims that attentional effects on emotional responses are independent of affective priming. Moreover, the effect was replicated even when target type was mixed from trial to trial (as opposed to being blocked), indicating that the effect is based on the immediate task relevance of stimuli and does not require an estab- lished mental set. Second, the pattern of data in Experiment 2 indicates clearly that the affective devaluation of distracting stimuli can be

generalized to other novel stimuli of the same general category. This

finding is consistent with the categorical basis of attentional selection needed to perform the localization task. Participants were required to

540 VOL. 14, NO. 6, NOVEMBER 2003




Jane E. Raymond, Mark J. Fenske, and Nader T. Tavassoli

Fig. 3. Average ratings of the emotional tone of images in Experiments 1 and 2. Results are shown

separately for the two attention conditions and, for Experiment 2, for trials on which the evaluated stimulus was the same token as the target in the prior localization task or a different token but the same type. Vertical bars indicate ± 1 SE.

ignore a pattern type, and so it follows that the effect of attention on

subsequent evaluation was also based on type.

GENERAL DISCUSSION

Our demonstration that prior attentional relevance has ramifica- tions for subsequent emotional evaluation complements existing evidence that emotional processes influence attention and indicates that the relation between attention and emotion is truly reciprocal. Some- what surprisingly, it is selective ignoring, as opposed to attending, that

appears to exert influence on emotional evaluation. The attentional

process of ignoring is known to involve suppression of perceptual processing of distracting objects (Moran & Desimone, 1985) and is

thought to prevent responses from being controlled by representations of task-irrelevant objects (Tipper, Brehaut, & Driver, 1990). Our results suggest that this top-down control over perceptual processing is extended to include emotional processing, even when the attention-

demanding task does not require emotional evaluation of stimuli. We have found an effect of ignoring that persists even when previously ignored stimuli become the subject of attention. This effect echoes the

findings of numerous negative priming studies in which perceptual re-

sponses to previously ignored objects were found to be slower than

perceptual responses to similar, previously attended objects (Tipper &

Cranston, 1985). The persistent effect of ignoring on emotional evaluation of stimuli may serve to reduce the emotional salience of distract-

ing stimuli so that, in future, they will be less likely to attract attention. In this way, attention and emotion may work together to prioritize processing. Although the neural mechanisms for attentional modulation of emotion remain unclear, they are likely to be cortical in origin, because recent human neuroimaging research has shown that attention

modulates activity in response to emotional facial expressions in cortical but not in subcortical emotion areas (Armony & Dolan, 2002; Vuilleumier et al., 2001).

The deleterious effect of ignoring a single brief presentation of an otherwise novel pattern may seem surprising in light of hundreds of

mere-exposure studies in which an unexpectedly positive emotional evaluation of stimuli has been found after repeated passive exposure to those or similar stimuli (Bornstein, 1989; Gordon & Holyoak, 1983; Reber et al., 2000; Zajonc, 2001). However, because previous studies did not manipulate attention to stimuli during exposure and the poten- tial consequences of active ignoring were overlooked, these studies failed to find what we discovered.

Mere-exposure effects have been explained in two different ways, neither of which predicts our results nor incorporates attentional processes. One view is that passive, low-level perceptual-encoding experience facilitates subsequent perception, causing feelings of familiarity to be interpreted as preference (Bornstein & D'Agostino, 1994; Sea-

mon, Brody, & Kauf, 1983). This view predicts that any prior experience with a stimulus, distracting or not, should enhance evaluations of it: Clearly our results do not support this view. The other account is

that preference results from classical conditioning of passively viewed stimuli to the absence of noxious events (Zajonc, 2001). This view also predicts that brief exposure to mildly distracting, but not otherwise unpleasant, stimuli should lead to preference, not devaluation,

particularly if the distracting information is successfully ignored. A

conditioning explanation applied to our experiment also predicts modulation by affective primes in our experimental trials, an effect we did not find. Clearly our results run counter to both views of the mere-

exposure effect and underscore the importance of considering attention in understanding the link between perceptual experience and emo-

VOL. 14, NO. 6, NOVEMBER 2003 541





tional appraisal. Although the phenomenon of mere exposure has been embraced by the advertising industry as justification for intrusive image presentations (Bartle, 1997; e.g., banner ads on a Web site), our results suggest that images presented while viewers are engaged in other attention-demanding tasks may engender negative rather than positive affect.

Acknowledgments - We thank Margaret Jackson and Gary Hill for their assistance in data collection. This research was supported by the Economic and Social Research Council (UK; R022250212) and the Center for eBusi- ness@MIT.

REFERENCES

Anderson, A.K., & Phelps, E.A. (2001). Lesions of the human amygdala impair enhanced perception of emotionally salient events. Nature, 411, 305-309.

Armony, J.L., & Dolan, R.J. (2002). Modulation of spatial attention by fear conditioned stimuli: An event-related fMRI study. Neuropsychologia, 40, 817-826.

Bartle, J. (1997). The advertising contribution. In L. Butterfield (Ed.), Excellence in advertising: The IPA guide to best practice (pp. 25-41). Oxford, England: Butterworth Heineman.

Bornstein, R.F. (1989). Exposure and affect: Overview and meta-analysis of research, 1968-1987. Psychological Bulletin, 106, 265-289.

Bornstein, R.F., & D'Agostino, PR. (1994). The attribution and discounting of perceptual fluency: Preliminary tests of a perceptual fluency/attributional model of the mere exposure effect. Social Cognition, 9, 103-128.

Bush, G., Luu, P., & Posner, M.I. (2000). Cognitive and emotional influences in anterior cingulate cortex. Trends in Cognitive Sciences, 4, 215-222.

Cohen, J.D., MacWhinney, B., Flatt, M., & Provost, J. (1993). PsyScope: A new graphic interactive environment for designing psychology experiments. Behavior Research Methods, Instruments, & Computers, 25, 251 -21 \.

Duncan, J. (1984). Selective attention and the organization of visual information. Journal of Experimental Psychology: General, 1 13, 501-517.

Eastwood, J.D., Smilek, D., & Menkle, P.M. (2001). Differential attentional guidance by unattended faces expressing positive and negative emotion. Perception & Psycho- physics, 63, 1004-1013.

Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening stimuli draw or hold

visual attention in subclinical anxiety? Journal of Experimental Psychology: Gen- eral, 130, 681-700.

Gordon, P., & Holyoak, K. (1983). Implicit learning and generalization of the "mere exposure" effect. Journal of Personality and Social Psychology, 3, 492-500.

Hopfinger, J.B., Buonocore, M.H., & Mangun, G.R. (2000). The neural mechanisms of top-down attentional control. Nature Neuroscience, 3, 284-291 .

Kastner, S., & Ungerleider, L.G. (2000). Mechanisms of visual attention in the human cortex. Annual Review of Neuroscience, 23, 315-341.

Le Doux, J.E. (1996). The emotional brain. New York: Simon & Schuster. Moran, J., & Desimone, R. (1985). Selective attention gates visual processing in the ex-

trastriate cortex. Science, 229, 782-784. Murphy, S.T., & Zajonc, R.B. (1993). Affect, cognition, and awareness: Affective priming

with optimal and suboptimal stimulus exposures. Journal of Personality and Social Psychology, 64, 723-739.

Ortony, A., Clore, G.L., & Collins, A. (1988). The cognitive structure of emotions. New York: Cambridge University Press.

Panksepp, J. (1998). Affective neuroscience. Oxford, England: Oxford University Press. Raymond, J.E., Shapiro, K.L., & Arnell, K.M. (1992). Temporary suppression of visual

processing in an RSVP task: An attentional blink? Journal of Experimental Psy- chology: Human Perception and Performance, 18, 849-860.

Reber, R., Winkielman, P., & Schwarz, N. (2000). Effects of perceptual fluency on affective judgments. Psychological Science, 9, 45^48.

Rolls, E.T. (1999). The brain and emotion. Oxford, England: Oxford University Press. Seamon, J.G., Brody, N., & Kauf, D.M. (1983). Affective discrimination of stimuli that are

not recognized: I. Effects of shadowing, masking, and cerebral laterality. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 544-555.

Tipper, S.P., Brehaut, J.C., & Driver, J. (1990). Selection of moving and static objects for the control of spatially directed action. Journal of Experimental Psychology: Hu- man Perception and Performance, 16, 492-504.

Tipper, S.P., & Cranston, M. (1985). Selective attention and pnming: Inhibitory and tacili- tatory effects of ignored primes. Quarterly Journal of Experimental Psychology, 37A, 591-611.

Tipper, S.P., & Driver, J. (1988). Negative priming between pictures and words in a selective attention task: Evidence for semantic processing of ignored stimuli. Memory & Cognition, 16, 64-70.

Vuilleumier, P., Armony, J.L., Driver, J., & Dolan, R.J. (2001). Effects of attention and emotion on face processing in the human brain: An event-related study. Neuron, 30, 1-20.

Vuilleumier, P., & Schwartz, S. (2001). Emotional facial expressions capture attention. Neurology, 56, 153-158.

Zajonc, R.B. (2001). Mere exposure: A gateway to the subliminal. Current Directions in Psychological Science, 10, 224-228.

(Received 7/18/02; Revision accepted 1 1/26/02)

542 VOL. 14, NO. 6, NOVEMBER 2003



selective attention determines emotional responses to novel visual stimuli

Documents